modin_pandas_drop_in_replacement_with_ray_engine.py

1import modin.pandas as pd
2import numpy as np
3import os
4
5# Optional: Force Modin to use a specific engine (Ray, Dask, or Unidist)
6# If not set, Modin will detect what is installed, defaulting to Ray.
7os.environ["MODIN_ENGINE"] = "ray" 
8
9# Create a sample dataset
10data = np.random.randint(0, 100, size=(2**10, 2**8))
11df = pd.DataFrame(data)
12
13# Modin API is identical to Pandas
14print("DataFrame Head:")
15print(df.head())
16
17# Perform a standard operation (e.g., mean)
18# This will be distributed across all available CPU cores
19mean_values = df.mean()
20print("\nColumn Means (first 5):")
21print(mean_values.head())
22
23# Verify the type is a Modin DataFrame
24print(f"\nObject type: {type(df)}")